In prompt driven development, you will achieve higher quality code output. If you provide the model a holistic product context for the goal you were trying to accomplish.

You need to act like you're the entire product team when you're prompting LLMs to write code for you. This is how you get better output. It's all about context, it's all about context, it's all about information that you give the model. Think of the model as a human, as a developer. The developer needs context. If the developer's gonna write high quality code, if the developer's gonna build the feature that you actually want them to build and have it work the way you want it to work and solve the user problem and achieve the business goal, you need to give them all that context. So when you're writing prompts into Cursor, Windsurf, Bolt, Lovable, V0, any of the other LLM native IDEs, think about this. Be the product manager, be the engineer, be the designer, be the QA engineer. All of those different roles when defining a scope of work and software development provide input. And think about the types of input that they provide into a scope of work and give that input into the prompt, put it into the prompt. That gets you a high quality prompt. That's the amount of context, that's the variety of context that you wanna provide to the model so that it can actually do the job that you want it to do. Because if you just tell the model to build you a feature that does X without having a user story, without having design specs, without having a technical approach and how to test it, how to know that it's working properly, it's gonna fill in all those gaps. It's gonna guess and it's gonna hallucinate, right? So help the model out, give it a lot of context, you'll have better outcomes. Peace.